Abstract

Stress testing has gained importance in financial institutions with the introduction of Basel II. Although discussed from many perspectives, the predominant use for stress testing is in predicting how a portfolio would respond to changes in the macroeconomic environment. The future environment is encapsulated in a macroeconomic scenario for an extreme situation and then fed through a scenario-based forecasting model. Validating stress testing models is inherently difficult, because financial institutions do not have historical data representing portfolio performance through many severe recessions. Data availability presents challenges for standard in-sample/out-of-sample tests. This chapter discusses these limitations and describes a suite of tests that may be employed to determine the robustness of stress test models. Particular emphasis is given to retail portfolios, which have received very little attention in the literature.

1. Why Stress Test?

In many fields, stress testing is a routine part of the job. Architecture and engineering have made extensive use of stress testing for decades, with some very interesting case examples such as the Citibank Tower in Manhattan (Morgenstern, 1995). Although those fields have rich traditions and well-developed scientific literature on the subject, sadly very little of it is applicable to banking.

Just as a structural engineer would want to run simulations of extreme winds on a sky scraper, bankers need to simulate the impact of extreme macroeconomic environments on their portfolios. Therefore, a stress test model must contain explicit...